Physical activity (PA) during pregnancy lowers the risk of complications during and after pregnancy, such as high blood pressure, gestational diabetes, and excessive gestational weight gain (Committee on Obstetric Practice, 2015; Evenson et al., 2014; Mottola et al., 2018; Moyer et al., 2016; Perales & Artal, 2017). The World Health Organization recommends that pregnant women be physically active at moderate intensity for at least 150 min/week and reduce their sedentary time (SED; Bull et al., 2020). Guidelines should be based on valid and reliable measurements of PA and SED, but estimates of PA and SED vary between different measurement tools (Fazzi et al., 2017; Strath et al., 2013). Considerations for the selection of a measurement tool include the purpose (e.g., surveillance and assessment of the effectiveness of an intervention), the variables of interest (e.g., total activity levels and energy expenditure), and practical factors (e.g., numbers of people being measured and cost; Strath et al., 2013).
Activity trackers offer objective measures of PA and SED via accelerometers and physiological sensors (e.g., heart rate; Wright et al., 2017). Activity trackers can estimate energy expenditure as well as PA duration, frequency, and intensity (e.g., moderate-to-vigorous intensity PA; Düking et al., 2018; Wright et al., 2017). Additionally, some activity trackers can recognize different types of activities, for example, walking, running, or cycling (Wright et al., 2017). Many consumer activity trackers are placed on the wrist and can be linked to an app or a platform to review the acquired data, assess data quality, and plan and deliver an intervention; in addition, data can be extracted for analysis (Wright et al., 2017). From a research perspective, there are a number of limitations associated with consumer activity trackers, including proprietary software and algorithms, and unknown details of software updates (Shei et al., 2022; Wright et al., 2017). While previous research studies have demonstrated relatively good accuracy of activity trackers in estimating total energy expenditure (TEE; O’Driscoll et al., 2020), there is a lack of rigorous and standardized validation of these devices in free-living settings (Argent et al., 2022; Düking et al., 2018). Also, no previous studies have validated the performance of activity trackers against the doubly labeled water (DLW) technique in pregnant women (Sattler et al., 2018).
DLW is expensive but is considered the “gold standard” for measuring TEE in free-living humans. The technique is based on the principle that the disappearance rate of the heavier stable isotope of hydrogen (H2) reflects the water turnover rate and the disappearance rate of the heavier stable isotope oxygen (18O) reflects both water and carbon dioxide (CO2) turnover rates. Therefore, with time, the difference between the disappearance rates of H2 and 18O represents the rate of CO2 production. Based on the energy equivalent of CO2, TEE can be estimated by the rate of CO2 production. The energy expended due to PA (PAEE) can be calculated from TEE (Westerterp et al., 2013) by subtracting the basal metabolic rate and the thermic effect of food.
Self-report questionnaires on PA are convenient and inexpensive and have shown acceptable reliability but low validity against accelerometers, pedometers, and diaries for pregnant women (Sattler et al., 2018; Schuster, 2016). The Pregnancy Physical Activity Questionnaire (PPAQ) was developed to assess PA specifically during pregnancy. PPAQ is a self-administrated, semiquantitative questionnaire with 32 questions designed to measure the duration and intensity (sedentary, light, moderate, and vigorous) of activities from various domains (household, occupation, transportation, and sport; Chasan-Taber et al., 2004). The Danish version of PPAQ (PPAQ-DK) has demonstrated acceptable reliability for measuring PA among pregnant women (Krøner et al., 2020), and the validity of PPAQ to measure total PA and SED using wrist-worn Actigraph GT3XP-BTLE (ActiGraph) and a wearable camera is currently being investigated by the author of the original tool (Grantome, n.d.). However, relying on only questionnaires, such as PPAQ, to measure the effectiveness of an intervention is not recommended (Sattler et al., 2018; Schuster, 2016). PPAQ has been validated against accelerometers (Chandonnet et al., 2012) and pedometers (Çırak et al., 2015) but estimates of PAEE, moderate-to-vigorous intensity PA, and SED obtained by PPAQ have not previously been compared with a consumer activity tracker or validated against DLW (Krøner et al., 2020).
The FitMum study aims to explore strategies to increase PA during pregnancy in women with low PA, and assess the health effects of PA (Knudsen et al., 2022; Roland et al., 2021). The primary outcome was moderate-to-vigorous intensity PA as measured by a Garmin Vivosport activity tracker. In addition, PPAQ-DK was used to assess PA practices among FitMum women. However, none of the tools have been validated against a criterion method in pregnancy. Thus, the present study aimed to validate TEE, PAEE, and PA levels using the Garmin Vivosport activity tracker and PAEE from PPAQ-DK against the DLW technique as the criterion method in pregnant women from the FitMum study. Moreover, we compared PAEE, moderate-to-vigorous intensity PA, and SED measures from the activity tracker and PPAQ-DK.
Methods
Setting, Participants, and Study Design
The study was part of the FitMum randomized controlled trial conducted at the Department of Gynaecology and Obstetrics, Copenhagen University Hospital—North Zealand, Hillerød, in the Capital Region of Denmark (Roland et al., 2021). Two hundred and twenty healthy, inactive (< 60 min/week of structured moderate-to-vigorous intensity PA) pregnant women were enrolled before gestational Week 15 (Visit 1). The first participant was included on October 1, 2018 and the last participant gave birth at the end of May 2021. After a 1-week baseline period, participants were randomized to one of two different PA interventions or standard care throughout pregnancy. In the 29th and the 35th gestational weeks, Visits 2 and 3 took place. Figure 1 presents the methods used and time points of the data collection.
—Methods used and time points of the data collection. DLW = doubly labeled water; PPAQ-DK = Danish pregnancy physical activity questionnaire; GA = gestational age.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Methods used and time points of the data collection. DLW = doubly labeled water; PPAQ-DK = Danish pregnancy physical activity questionnaire; GA = gestational age.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Methods used and time points of the data collection. DLW = doubly labeled water; PPAQ-DK = Danish pregnancy physical activity questionnaire; GA = gestational age.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
Doubly Labeled Water
DLW was administered at Visit 2 to 134 participants. The DLW test was made available to interested participants until we ran out of DLW supply (due to economic reasons). Two baseline urine samples were collected for the determination of background enrichment. The DLW (Sercon Limited) was obtained from two larger batches containing D2O (99.8% enrichment) and 18O (10% enrichment). The participant’s body weight was measured and the participant was administered a glass of water for oral intake of 0.1 g of 99.98 % D2O and 1.6 g of 10% 18O per kg body weight. The exact date and time of DLW intake were recorded. Aliquot doses of DLW were saved at −80 °C as reference samples. During the following 2 weeks, the participant collected a single urine sample of at least 10 ml in the morning (not the first void) after 1, 4, 7, 11, and 14 days. Each participant recorded the exact date and time of the void. Participant compliance was excellent; only five participants missed one urine collection and two participants missed two or three collections. The participants were instructed to keep the samples at −20 °C after collection, and the samples were transported to Copenhagen University Hospital—North Zealand as soon as possible after Day 14 for storage at −80 °C. Analysis of the samples was performed at Clinical Metabolomics Core Facility, Copenhagen University Hospital—Rigshospitalet.
2/1H2 and C18/16O2 were determined by isotope ratio mass spectrometry using a Thermo Delta V Advantage continuous-flow isotope ratio mass spectrometer system equipped with a Thermo GasBench II (Thermo Scientific). All stable 2/1H and 18/16O isotope ratio measurements were expressed in d per mil unit (o/oo) versus the international reference materials Standard Mean Ocean Water and Standard Light Antarctic Precipitation (IAEA). Moreover, for each batch of urine analysis, a six-point water calibration curve was determined with known 2/1H218/16O enrichments. Calibrators and urine samples were then prepared for 2/1H determination. After uncapping a 12-ml exetainer (Labco International), 5 mg of activated charcoal (Fisher Scientific) was introduced into the exetainer followed by a platinum catalytic rod (Thermo Scientific). The activated charcoal was added to remove any potential contaminants in the urine sample that might poison the catalyst. After putting 0.2 ml of urine into the exetainer, the exetainer was recapped and placed into the GasBench II and flushed with 2% H2 in helium (99.999% H2 and 99.996% helium) for 7 min. The samples were equilibrated for at least 4 hr with the H2 at room temperature. After equilibration, six aliquots of headspace were injected into the isotope ratio mass spectrometry for 2/1H isotope ratio measurement against the reference H2O Vienna Standard Mean Ocean Water and Standard Light Antarctic Precipitation (VSMOV2). For C18/16O2 analysis, 0.2 ml of calibrators or urine was added to the 12-ml exetainer, and the exetainer was recapped and placed into the GasBench II and flushed with 0.9% CO2 in helium (99.999% CO2 and 99.996% helium) for 7 min. The samples were equilibrated for 24 hr with the CO2 at room temperature after which six aliquots of the headspace were injected into the isotope ratio mass spectrometry for 18/16O isotope ratio measurement against the reference H2O Standard Light Antarctic Precipitation 2 (SLAP2). No samples were lost during the analysis.
The slope-intercept protocol was used to calculate TEE (Bhutani, 2015). First, the enrichment estimate rates of H2 and 18O were calculated. Each sample data (enrichment estimate) were baseline corrected. Then, the dilution space for the two isotopes (oxygen [kO] and deuterium [kH]) was calculated by using the natural logarithm of the mean enrichment for both H2 and 18O with the number of days elapsed. The total body water was derived from the average dilution space of H2 and 18O and then divided by the factor for correction for in vivo isotopic exchange, which is 1.04 for H2 and 1.01 for 18O. The ratio between H2 and 18O dilution spaces was (mean ± SD) 1.02 ± 0.05 (Bhutani, 2015). The rate of CO2 production (rCO2) was calculated as (total body water/2; kO—kH). Subsequently, TEE was calculated, using the modified Weir equation, as 22.4 (3.9 [rCO2/Food Quotient] + 1.1 [rCO2]) × 4.184/1,000, with the Food Quotient assumed to be 0.85 (Löf, 2011; Speakman et al., 2021). Basal metabolic rate was estimated from body weight, height, and age by an equation derived from the Harris–Benedict equation for pregnant women (Hronek et al., 2009). Finally, PAEE was determined by subtraction of the basal metabolic rate and the thermic effect of food (assumed to be 10% of TEE) from TEE (Hronek et al., 2009; Westerterp et al., 2013). PA level was calculated by dividing TEE by the basal metabolic rate (Löf, 2011).
Activity Tracker
At Visit 1, participants were provided with a wrist-worn consumer activity tracker with a built-in heart rate monitor, known as photoplethysmography, and an accelerometer (Garmin Vivosport, Garmin International; Garmin International, n.d.) which had to be worn on the nondominant wrist 24/7 throughout pregnancy. The Garmin Vivosport is lightweight, has a battery life of up to 7 days, and can store activity and heart rate data for up to 14 days between synchronization (Garmin International, n.d.). The research team monitored compliance with wearing the tracker and syncing data through the Fitabase research platform (Fitabase). The participant was instructed on how to synchronize and charge the tracker regularly; an email reminder was sent if a participant was not syncing for more than 7 days. To be included in the data analysis, a valid day of measurement comprised at least 12 hr of daily wear time (from 6 a.m. to 12 a.m.) with at least four valid days (weekdays and/or weekend days) per week. Validation and inclusion of days and weeks were based on heart rate data which were sampled every 15 s by the activity tracker.
Activity estimates were based on Garmin’s proprietary algorithms which were unavailable to the researchers. The participant’s height, body weight, age, and sex were entered in the Garmin Connect app at inclusion; body weight was measured at Visits 2 and 3 and reentered into the app. Basal metabolic rate, PAEE, and moderate-to-vigorous intensity PA were derived directly from the activity tracker. Only PA with a metabolic equivalent of task value of ≥ 3 in bouts of at least 10 consecutive minutes was recorded as moderate-to-vigorous intensity PA by the tracker. We calculated TEE by adding PAEE, basal metabolic rate, and thermic effect of food which was assumed to be 10% of TEE (Hronek et al., 2009; Westerterp et al., 2013). The tracker categorized time into sedentary, active, or highly active. Total sedentary time was defined as little to no activity monitored, such as minimal movement, sitting, resting, or sleeping (Garmin International, n.d.). The activity tracker also obtained total sleep time, and we calculated sedentary awake time (i.e., SED) by subtracting sleep time from total sedentary time.
Pregnancy PA Questionnaire
The PPAQ is a subjective instrument estimating PA in the current trimester (Chasan-Taber et al., 2004). Adapted from the international PA questionnaire (Craig et al., 2003), the PPAQ collects data on different activities, for example, household, occupational, and sports. The FitMum research group translated PPAQ into Danish (PPAQ-DK) and validated it in a Danish pregnant population (Krøner et al., 2020). The PPAQ-DK was electronically distributed to the participants immediately after Visits 1, 2, and 3. The PPAQ-DK has questions about time spent in sedentary activities (n = 5; Barone Gibbs et al., 2020), light-intensity activities (n = 8), moderate-intensity activities (n = 15), and vigorous-intensity activities (n = 2), as well as two open questions if some activities were not stated. PAEE was calculated as the number of minutes spent in each reported activity multiplied by its metabolic equivalence of task value (Cohen et al., 2013). To calculate moderate-to-vigorous intensity PA, minutes at moderate intensity and vigorous intensity were added.
Comparisons of the Three Methods
DLW in a 14-day period after Visit 2 was used as a criterion method for TEE, PAEE, and PA level. For comparison, the averaged tracker variables during the same 14-day period and PPAQ-DK answered at Visit 2 were used. When comparing PAEE, moderate-to-vigorous intensity PA; SED from activity tracker data with PPAQ-DK at Visits 1, 2, and 3; and the tracker data were averaged from Visit 1 to randomization (six full days), from randomization to Visit 2, and from Visit 2 to Visit 3, respectively. Measurements from the activity tracker and PPAQ were used for comparison purposes.
Statistics
All data were analyzed using R (version 4.0.4, 2021-02-15; R Core Team, [2020]). Descriptive statistics were presented as mean (SD) or median (interquartile range). Pearson correlation was used to assess the relationship between estimates of TEE, PAEE, and PA level, respectively, from the activity tracker and DLW, and the relationship of PAEE from PPAQ-DK and DLW. Also, Pearson correlation was used to assess the relationship between PAEE, moderate-to-vigorous intensity PA, and SED estimates, respectively, from the activity tracker and PPAQ-DK at Visits 1, 2, and 3. The agreement between methods was based on the correlation coefficient (r) classified as weak (.10–.39), moderate (.40–.69), strong (.70–.89), or very strong (.90–1.00; Schober & Schwarte, 2018). Bland–Altman analysis (Bland & Altman, 1999) was used to explore the levels of agreement for TEE, PAEE, and PA levels, respectively, between the activity tracker and DLW and the agreement of PAEE between PPAQ-DK and DLW. Additionally, Bland–Altman analysis was used to assess the agreement of PAEE, moderate-to-vigorous intensity PA, and SED, respectively, between the activity tracker and PPAQ-DK at Visits 1, 2, and 3. In all the Bland–Altman plots, the means of the two methods were plotted on the x-axis against the difference in the y-axis, except for DLW as this was considered the criterion method and hence directly plotted in the x-axis (Hills et al., 2014). The mean absolute percentage error was calculated for the tracker and PPAQ compared to DLW (Argent et al., 2022; Johnston et al., 2021). Finally, linear regression analysis was performed to find a possible variation between the methods (i.e., proportional bias) presented as slope (95% confidence interval). The dependent and independent variables were the same as in the Bland–Altman plots. The significance level was set at 5%.
Results
At enrollment, participants were 31.5 ± 4.3 years old, had a gestational age of 12.9 (9.4–13.9) weeks, a body weight of 75.4 ± 15.3 kg, and a median prepregnancy body mass index of 24.1 (21.8–28.7) kg/m2. During the 6-day baseline period, participants wore the activity tracker for a total of 1,278 days out of 1,314 potential days (97%; 6 [3–6 days]). From randomization to delivery, participants wore the tracker for a total of 32,421 days out of 42,041 potential days (77%; 172 [4–230 days]). For the comparison of the activity tracker and DLW data, 134 participants were included, whereas 133 participants were included in the comparison of PPAQ-DK and DLW data. There were no significant differences in baseline characteristics (body weight, age, educational level, gestational age, and parity) for those who had DLW and those who did not. The number of participants included in the comparison of data from PPAQ-DK and activity trackers was 218 at Visit 1, 181 at Visit 2, and 165 at Visit 3.
Validity of Activity Tracker and PPAQ-DK Against DLW
Summary statistics of the method comparisons are shown in Table 1. A moderate correlation was found between TEE from the activity tracker and DLW (r = .63; p < .001; Figure 2A). However, PAEE and PA levels from the activity tracker did not correlate with PAEE and PA levels from DLW (Figure 2B and 2C). Moreover, PAEE from PPAQ-DK did not correlate with PAEE from DLW (Figure 2D). The activity tracker overestimated TEE (mean bias: 503 kcal/day; Figure 3A), and the proportional bias was not significant (slope = −0.01; [−0.2 to 0.2]; p = .919; Table 1). Also, the activity tracker overestimated PAEE (mean bias: 303 kcal/day; Figure 3B) and PA level (mean bias: 0.2; Figure 3C) when compared with DLW. For PAEE around 400 kcal/day and PA level around 1.3, the activity tracker shifted gradually to underestimate values compared to DLW. The thermic effect of food is not always included in calculations of TEE and PAEE (Hallal et al., 2013), that is, TEE = BMR + PAEE + thermic effect of food or TEE = BMR + PAEE. Notably, when we omitted the thermic effect of food in our calculations, we saw a relatively good agreement for TEE (mean bias: 222 kcal/day) and PAEE (mean bias: 72 kcal/day), respectively, between the activity tracker and DLW. PPAQ-DK overestimated PAEE (mean bias: 1,513 kcal/day) compared with DLW (Figure 3D) with insignificant proportional bias (slope = −0.8; [−1.6 to 0.2]; p = .050; Table 1). When comparing data from the tracker and DLW, mean absolute percentage error was 18.6% (11.7%) for TEE, 143.0% (213.8%) for PAEE, and 11.8% (9.3%) for PA level. When comparing data from PPAQ and DLW, mean absolute percentage error was 241.0% (172.6%) for PAEE.
Summary Statistics of the Methods’ Comparisons
| Variable | Contrast | Time | n | Mean bias | 95% LOA | Proportional bias, intercept (slope) | [95% CI] | p | |
|---|---|---|---|---|---|---|---|---|---|
| Total energy expenditure (kcal/day) | Tracker vs. DLW | Visit 2 (29th GA week) | 134 | 503 | −133 | 1,139 | 477 (−0.01) | [−0.2, 0.2] | .919 |
| Physical activity energy expenditure (kcal/day) | 134 | 303 | −284 | 890 | 694 (−1) | [−1.4, −0.8] | <.000 | ||
| Physical activity level (TEE/BMR) | 134 | 0.2 | −0.2 | 0.5 | 1.6 (−1) | [−1.3, −0.8] | <.000 | ||
| Physical activity energy expenditure (kcal/day) | PPAQ-DK vs. DLW | 133 | 1,513 | 308 | 2,718 | 1,786 (−0.8) | [−1.6, 0.2] | .050 | |
| Physical activity energy expenditure (kcal/day) | Tracker vs. PPAQ-DK | Visit 1 (GA ≤ 15 weeks) | 218 | −1,286 | −2,565 | −6 | 390 (−1.5) | [−1.6, −1.4] | <.000 |
| Visit 2 (29th GA week) | 179 | −1,356 | −2,740 | 28 | 550 (−1.6) | [−1.7, −1.4] | <.000 | ||
| Visit 3 (35th GA week) | 160 | −1,046 | −2,570 | 478 | 670 (−1.4) | [−1.6, −1.3] | <.000 | ||
| Moderate-to-vigorous intensity physical activity (min/day) | Tracker vs. PPAQ-DK | Visit 1 (GA ≤ 15 weeks) | 218 | −90 | −230 | 50 | 5 (−1.9) | [−2, −1.8] | <.000 |
| Visit 2 (29th GA week) | 181 | −86 | −219 | 47 | 12 (−1.9) | [−2, −1.8] | <.000 | ||
| Visit 3 (35th GA week) | 165 | −72 | −222 | 78 | 11 (−1.8) | [−1.9, −1.7] | <.000 | ||
| Sedentary time (hr/day) | Tracker vs. PPAQ-DK | Visit 1 (GA ≤ 15 weeks) | 218 | 6.8 | 1.3 | 12.3 | 18 (−1.2) | [−1.4, −1.05] | <.000 |
| Visit 2 (29th GA week) | 181 | 7.2 | 2.2 | 12.2 | 21 (−1.5) | [−1.6, −1.3] | <.000 | ||
| Visit 3 (35th GA week) | 162 | 8.1 | 2.2 | 13.8 | 18 (−1.1) | [−1.4, −0.9] | <.000 | ||
Note. TEE = total energy expenditure; BMR = basic metabolic rate; DLW = doubly labeled water; PPAQ-DK = Danish pregnancy physical activity questionnaire; GA = gestational age; n = number; LOA = limit of agreement; CI = confidence interval.
—Correlations between the activity tracker and PPAQ-DK outcomes, respectively, and DLW outcomes. (A) TEE: activity tracker versus DLW; (B) PAEE: activity tracker versus DLW; (C) PAL: activity tracker versus DLW; and (D) PAEE: PPAQ-DK versus DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Correlations between the activity tracker and PPAQ-DK outcomes, respectively, and DLW outcomes. (A) TEE: activity tracker versus DLW; (B) PAEE: activity tracker versus DLW; (C) PAL: activity tracker versus DLW; and (D) PAEE: PPAQ-DK versus DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Correlations between the activity tracker and PPAQ-DK outcomes, respectively, and DLW outcomes. (A) TEE: activity tracker versus DLW; (B) PAEE: activity tracker versus DLW; (C) PAL: activity tracker versus DLW; and (D) PAEE: PPAQ-DK versus DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes, respectively, against DLW outcomes: (A) differences between activity tracker and DLW; (B) PAEE: differences between activity tracker and PPAQ-DK versus DLW; (C) PAL: differences between activity tracker and DLW; and (D) PAEE: differences between PPAQ-DK and DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes, respectively, against DLW outcomes: (A) differences between activity tracker and DLW; (B) PAEE: differences between activity tracker and PPAQ-DK versus DLW; (C) PAL: differences between activity tracker and DLW; and (D) PAEE: differences between PPAQ-DK and DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes, respectively, against DLW outcomes: (A) differences between activity tracker and DLW; (B) PAEE: differences between activity tracker and PPAQ-DK versus DLW; (C) PAL: differences between activity tracker and DLW; and (D) PAEE: differences between PPAQ-DK and DLW. PPAQ-DK = Danish pregnancy physical activity questionnaire; TEE = total energy expenditure; DLW = doubly labeled water; PAEE = PA energy expenditure; PAL = PA level; BMR = basic metabolic rate; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
Comparison Between Activity Tracker and PPAQ-DK
PAEE from the activity tracker and PPAQ-DK correlated weakly at Visits 1, 2, and 3 (r: .15–.22; Figure 4A–C). Moderate-to-vigorous intensity PA and SED from the activity tracker and PPAQ-DK correlated weakly (r: .15–.17; Figure 4F, G, and H) or not at all (Figures 4D, 4E, and 4I). The pattern of the comparisons between PAEE, moderate-to-vigorous intensity PA, and SED by the activity tracker and PPAQ-DK was consistent throughout the three visits (Figure 5, Table 1). The mean biases between PPAQ-DK and the activity tracker were −1,286, −1,356, and −1,046 kcal/day for PAEE (Figure 5A–C, Table 1); −90, −86, and −72 min/day for moderate-to-vigorous intensity PA (Figure 5D–F, Table 1); and 6.8, 7.2, and 8.1 hr/day for SED (Figure 5G–I, Table 1) at Visits 1, 2, and 3, respectively. The activity tracker consistently reported PAEE and moderate-to-vigorous intensity PA to be lower than reported by the PPAQ-DK. On the contrary, the activity tracker reported SED to be higher than reported by the PPAQ-DK. The linear regression analysis showed consistent and significant proportional biases (p < .000) between the activity tracker and PPAQ-DK for all variables and at all visits (Table 1).
—Correlations between activity tracker and PPAQ-DK outcomes. PAEE: activity tracker versus PPAQ-DK at Visits 1 (A), 2 (B), and 3 (C); MVPA: activity tracker versus PPAQ-DK at Visits 1 (D), 2 (E), and 3 (F); and SED: activity tracker versus PPAQ-DK at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Correlations between activity tracker and PPAQ-DK outcomes. PAEE: activity tracker versus PPAQ-DK at Visits 1 (A), 2 (B), and 3 (C); MVPA: activity tracker versus PPAQ-DK at Visits 1 (D), 2 (E), and 3 (F); and SED: activity tracker versus PPAQ-DK at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Correlations between activity tracker and PPAQ-DK outcomes. PAEE: activity tracker versus PPAQ-DK at Visits 1 (A), 2 (B), and 3 (C); MVPA: activity tracker versus PPAQ-DK at Visits 1 (D), 2 (E), and 3 (F); and SED: activity tracker versus PPAQ-DK at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes: PAEE: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (A), 2 (B), and 3 (C); MVPA: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (D), 2 (E), and 3 (F); and SED: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes: PAEE: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (A), 2 (B), and 3 (C); MVPA: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (D), 2 (E), and 3 (F); and SED: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
—Bland–Altman plots for the activity tracker and PPAQ-DK outcomes: PAEE: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (A), 2 (B), and 3 (C); MVPA: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (D), 2 (E), and 3 (F); and SED: differences between PPAQ-DK and activity tracker versus sums at Visits 1 (G), 2 (H), and 3 (I). PPAQ-DK = Danish pregnancy physical activity questionnaire; PAEE = PA energy expenditure; MVPA = moderate-to-vigorous intensity PA; SED = sedentary time; LOA = limit of agreement.
Citation: Journal for the Measurement of Physical Behaviour 6, 1; 10.1123/jmpb.2022-0033
Discussion
This is the first study to validate a consumer activity tracker and PPAQ-DK against DLW in pregnant women. TEE from the activity tracker and DLW correlated moderately well, which might be due to the dominance of BMR in both estimates. Also, we found better agreement of PAEE estimates from the activity tracker than from PPAQ-DK when compared with DLW. Moreover, throughout our study, pregnancy tracker estimates of PA were lower and estimates of SED were higher than estimates from PPAQ-DK. Participants showed good compliance with wearing the activity tracker during the baseline period (97%) and throughout pregnancy (77%), which is comparable to another study among pregnant women (Grym et al., 2019).
A previous review of the validity of different Garmin activity trackers (Evenson & Spade, 2020) concluded that the error for determining energy expenditure generally was considerable compared to a criterion method (e.g., indirect calorimetry and triaxial accelerometer). However, that review did not include the Garmin Vivosport activity tracker used in the present study. Wahl et al. tested the validity of 11 different activity trackers during exercise in 20 healthy sports students, including three Garmin trackers (but not Vivosport) and used indirect calorimetry as the criterion method (Basset et al., 2017). Using an exercise protocol with different velocities, they found that the activity trackers generally overestimated energy expenditure when exercise velocity was low and underestimated energy expenditure when PA intensity increased (Basset et al., 2017). Murakami et al. tested the validity of 12 activity trackers, including Garmin Vivofit, in 19 healthy women and men (Murakami et al., 2016, 2019). Participants wore trackers during a standardized day in a metabolic chamber as well as during 15 free-living days where PAEE was also determined by DLW. They found that most of the trackers underestimated PAEE, and Garmin Vivofit underestimated PAEE during standardized (mean bias: −499.3 kcal/day) and free-living conditions (mean bias: −727.8 kcal/day). In comparison, we found that Garmin Vivosport overestimated PAEE compared to DLW. Two different Fitbit activity trackers have also been validated against DLW (Shook et al., 2022; Siddall et al., 2019). The Fitbit Surge tended to underestimate TEE (mean bias: −656 kcal/day) in 20 military personnel (Siddall et al., 2019), whereas the Fitbit Alta HR showed moderate to strong agreement and correlation of TEE compared to DLW with small degrees of overestimation during two 14-day assessment periods (means bias: 17 and 76 kcal/day) in 24 healthy adults (Shook et al., 2022). Possible reasons for conflicting results between studies are different hardware and software in the multitude of commercial activity trackers on the market that are constantly updated (Henriksen et al., 2018; Woolley et al., 2019). Also, DLW is the gold standard for measuring free-living TEE, but the calculation of PAEE from DLW relies on several presumptions (Westerterp, 2017).
The PPAQ and PPAQ-DK have not previously been validated against DLW, but Besson et al. assessed the validity of the recent PA questionnaire against DLW in healthy adults (Besson et al., 2010). They found that the recent PA questionnaire underestimated PAEE which is in contrast to our findings for PPAQ-DK. Pedersen et al. tested another PA questionnaire, the PA Scale, against a combined accelerometer and heart rate monitor in Danish adults (Pedersen et al., 2018). They found that the PA Scale overestimated PAEE, which is similar to our findings. Brett et al. reported that PPAQ highly overestimated moderate-to-vigorous intensity PA when compared with the omniaxial Actical (Brett et al., 2015), which is in line with our results. Also, a systematic review of PA questionnaires for pregnant women found that the PPAQ validity was low in assessing moderate-to-vigorous intensity PA (Sattler et al., 2018). Similar to our findings, Barone Gibbs et al. found that PPAQ significantly underestimated SED when compared with both a thigh-worn activPAL3 micro (PAL Technologies Ltd.) (criterion method) and a waist-worn Actigraph GT3X (Barone Gibbs et al., 2020). Our findings affirmed that self-reported data tend to overestimate PA and underestimate SED (Chinapaw et al., 2009).
A strength of this study is the use of the DLW technique, the gold standard method for determining TEE under free-living conditions, in a high number of pregnant women. We are the first to validate both the activity tracker Garmin Vivosport and PPAQ-DK against DLW in pregnant women. A further strength is that the measurements from the activity tracker and PPAQ-DK were compared in all three trimesters of pregnancy. A limitation of the present study is that the activity estimates from the activity tracker were based on Garmin’s proprietary algorithms which were unavailable to the researchers. Moreover, the Garmin Vivosport software was automatically updated throughout the study period, which probably influenced the assessment of the PA metrics from the tracker.
Conclusions
PAEE estimates from the consumer activity tracker Garmin Vivosport were superior to estimates from PPAQ-DK when compared to DLW as the criterion method, but the absolute error of both the tracker and PPAQ-DK was significant. In addition, TEE from the tracker and DLW correlated moderately well. Moreover, PAEE and moderate-to-vigorous intensity PA measured by the activity tracker were lower throughout pregnancy and SED was higher than reported using PPAQ-DK. Thus, newer consumer activity trackers complement questionnaires to estimate energy expenditure, PA, and SED. In addition, activity trackers might motivate pregnant women to increase their PA levels.
Acknowledgments
The authors would like to acknowledge all women participating in FitMum. Special thanks to students and research assistants who were part of performing interventions and collecting data. Additionally, we would like to thank the technical staff at the Clinical Research Unit, Department of Clinical Research, Copenhagen University Hospital—North Zealand, Hillerød; especially Susanne Månsson and Charlotte Pietraszek who engaged themselves in planning practicalities and collecting data, as well as the staff at the Department of Gynaecology and Obstetrics, Copenhagen University Hospital—North Zealand, Hillerød. Funding: FitMum was funded by the Independent Research Fund Denmark (8020-00353B), TrygFonden (128509), Copenhagen Center for Health Technology (061017), Beckett-Fonden (17-2-0883), Aase and Ejnar Danielsens Fond (10-002052), and Familien Hede Nielsens Fond (2017-1142). In addition, funding was provided by the University of Copenhagen and Copenhagen University Hospital, North Zealand, Hillerød. Ethics Approval and Consent to Participate: The study was approved by the Danish National Committee on Health Research Ethics (#H-18011067) and the Danish Data Protection Agency (#P-2019-512). The study adheres to the principles of the Helsinki Declaration. Written informed consent was obtained at inclusion. Availability of Data and Materials: The Danish Data Protection Agency has not approved sharing of data.
References
Argent, R., Hetherington-Rauth, M., Stang, J., Tarp, J., Ortega, F.B., Molina-Garcia, P., Schumann, M., Bloch, W., Cheng, S., Grøntved, A., Brønd, J.C., Ekelund, U., Sardinha, L.B., & Caulfield, B. (2022). Recommendations for determining the validity of consumer wearables and smartphones for the estimation of energy expenditure: Expert statement and checklist of the INTERLIVE Network. Sports Medicine, 52(8), 1817–1832. https://doi.org/10.1007/s40279-022-01665-4
Barone Gibbs, B., Paley, J.L., Jones, M.A., Whitaker, K.M., Connolly, C.P., & Catov, J.M. (2020). Validity of self-reported and objectively measured sedentary behavior in pregnancy. BMC Pregnancy and Childbirth, 20(1), Article 99. https://doi.org/10.1186/s12884-020-2771-z
Basset, F.A., Passfield, L., Wahl, Y., Düking, P., Droszez, A., Wahl, P., & Mester, J. (2017). Criterion-validity of commercially available physical activity tracker to estimate step count, covered distance and energy expenditure during sports conditions. Frontiers in Physiology, 8, Article 725. https://doi.org/10.3389/fphys.2017.00725
Besson, H., Brage, S., Jakes, R.W., Ekelund, U., & Wareham, N.J. (2010). Estimating physical activity energy expenditure, sedentary time, and physical activity intensity by self-report in adults. The American Journal of Clinical Nutrition, 91(1), 106–114. https://doi.org/10.3945/AJCN.2009.28432
Bhutani, S. (2015). Special considerations for measuring energy expenditure with doubly labeled water under atypical conditions. Journal of Obesity & Weight Loss Therapy, 5(Suppl. 5), 1–20. https://doi.org/10.4172/2165-7904.s5-002
Bland, J.M., & Altman, D.G. (1999). Measuring agreement in method comparison studies [Article]. Statistical Methods in Medical Research, 8(2), 135–160. https://doi.org/10.1177/096228029900800204
Brett, K.E., Wilson, S., Ferraro, Z.M., & Adamo, K.B. (2015). Self-report pregnancy physical activity questionnaire overestimates physical activity. Canadian Journal of Public Health, 106(5), 297–302. https://doi.org/10.17269/CJPH.106.4938
Bull, F.C., Al-Ansari, S.S., Biddle, S., Borodulin, K., Buman, M.P., Cardon, G., Carty, C., Chaput, J.P., Chastin, S., Chou, R., Dempsey, P.C., Dipietro, L., Ekelund, U., Firth, J., Friedenreich, C.M., Garcia, L., Gichu, M., Jago, R., Katzmarzyk, P.T., … Willumsen, J.F. (2020). World Health Organization 2020 guidelines on physical activity and sedentary behaviour. British Journal of Sports Medicine, 54(24), 1451–1462. https://doi.org/10.1136/bjsports-2020-102955
Chandonnet, N., Saey, D., Alméras, N., & Marc, I. (2012). French pregnancy physical activity questionnaire compared with an accelerometer cut point to classify physical activity among pregnant obese women. PLoS One, 7(6), Article 38818. https://doi.org/10.1371/journal.pone.0038818
Chasan-Taber, L., Schmidt, M.D., Roberts, D.E., Hosmer, D., Markenson, G., & Freedson, P.S. (2004). Development and validation of a pregnancy physical activity questionnaire. Medicine & Science in Sports & Exercise, 36(10), 1750–1760. https://doi.org/10.1249/01.MSS.0000142303.49306.0D
Chinapaw, M.J., Slootmaker, S.M., Schuit, A.J., van Zuidam, M., & van Mechelen, W. (2009). Reliability and validity of the Activity Questionnaire for Adults and Adolescents (AQuAA). MC Medical Research Methodology, 9, Article 58. https://doi.org/10.1186/1471-2288-9-58
Çırak, Y., Yılmaz, G.D., Demir, Y.P., Dalkılınç, M., & Yaman, S. (2015). Pregnancy physical activity questionnaire (PPAQ): Reliability and validity of Turkish version. Journal of Physical Therapy Science, 27(12), 3703–3709. https://doi.org/10.1589/jpts.27.3703
Cohen, T.R., Plourde, H., & Koski, K.G. (2013). Use of the Pregnancy Physical Activity Questionnaire (PPAQ) to identify behaviours associated with appropriate gestational weight gain during pregnancy. Journal of Physical Activity & Health, 10(7), 1000–1007. https://doi.org/10.1123/jpah.10.7.1000
Committee on Obstetric Practice. (2015). Committee opinion: Physical activity and exercise during pregnancy and the postpartum period. The American College of Obstetricians and Gynecologists, 650, 1–8.
Craig, C.L., Marshall, A.L., Sjöström, M., Bauman, A.E., Booth, M.L., Ainsworth, B.E., Pratt, M., Ekelund, U., Yngve, A., Sallis, J.F., & Oja, P. (2003). International physical activity questionnaire: 12-Country reliability and validity. Medicine & Science in Sports & Exercise, 35(8), 1381–1395. https://doi.org/10.1249/01.MSS.0000078924.61453.FB
Düking, P., Fuss, F.K., Holmberg, H.C., & Sperlich, B. (2018). Recommendations for assessment of the reliability, sensitivity, and validity of data provided by wearable sensors designed for monitoring physical activity. JMIR MHealth and UHealth, 6(4), Article 9341. https://doi.org/10.2196/mhealth.9341
Evenson, K.R., Barakat, R., Brown, W.J., Dargent-Molina, P., Haruna, M., Mikkelsen, E.M., Mottola, M.F., Owe, K.M., Rousham, E.K., & Yeo, S. (2014). Guidelines for physical activity during pregnancy: Comparisons from around the world. American Journal of Lifestyle Medicine, 8(2), 102–121. https://doi.org/10.1177/1559827613498204
Evenson, K.R., & Spade, C.L. (2020). Review of validity and reliability of Garmin activity trackers. Journal for the Measurement of Physical Behaviour, 3(2), 170–185. https://doi.org/10.1123/jmpb.2019-0035
Fazzi, C., Saunders, D.H., Linton, K., Norman, J.E., & Reynolds, R.M. (2017). Sedentary behaviours during pregnancy: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 14(1), Article 32. https://doi.org/10.1186/s12966-017-0485-z
Garmin International. (n.d.). https://www.garmin.com
Grantome. (n.d.). Update and Novel Validation of the Pregnancy Physical Activity Questionnaire (PPAQ). Retrieved July 23, 2020, from https://grantome.com/grant/NIH/R21-HD094565-02
Grym, K., Niela-vilén, H., Ekholm, E., Hamari, L., Azimi, I., Rahmani, A., Liljeberg, P., Löyttyniemi, E., & Axelin, A. (2019). Feasibility of smart wristbands for continuous monitoring during pregnancy and one month after birth. BMC Pregnancy Childbirth, 19, Article 34. https://doi.org/10.1186/s12884-019-2187-9
Hallal, P.C., Reichert, F.F., Clark, V.L., Cordeira, K.L., Menezes, A.M.B., Eaton, S., Ekelund, U., & Wells, J.C. (2013). Energy expenditure compared to physical activity measured by accelerometry and self-report in adolescents: A validation study. PLoS One, 8(11), Article 77036. https://doi.org/10.1371/journal.pone.0077036
Henriksen, A., Haugen Mikalsen, M., Woldaregay, A.Z., Muzny, M., Hartvigsen, G., Hopstock, L.A., & Grimsgaard, S. (2018). Using fitness trackers and smartwatches to measure physical activity in research: Analysis of consumer wrist-worn wearables. Journal of Medical Internet Research, 20(3), Article 110. https://doi.org/10.2196/jmir.9157
Hills, A.P., Mokhtar, N., & Byrne, N.M. (2014). Assessment of physical activity and energy expenditure: An overview of objective measures. Frontiers in Nutrition, 1, Article 5. https://doi.org/10.3389/fnut.2014.00005
Hronek, M., Zadak, Z., Hrnciarikova, D., Hyspler, R., & Ticha, A. (2009). New equation for the prediction of resting energy expenditure during pregnancy. Nutrition, 25(9), 947–953. https://doi.org/10.1016/j.nut.2009.02.011
Johnston, W., Judice, P.B., Molina García, P., Mühlen, J.M., Lykke Skovgaard, E., Stang, J., Schumann, M., Cheng, S., Bloch, W., Brønd, J.C., Ekelund, U., Grøntved, A., Caulfield, B., Ortega, F.B., & Sardinha, L.B. (2021). Recommendations for determining the validity of consumer wearable and smartphone step count: Expert statement and checklist of the INTERLIVE network. British Journal of Sports Medicine, 55(14), 780–793. https://doi.org/10.1136/BJSPORTS-2020-103147
Knudsen, S.D.P., Alomairah, S.A., Roland, C.B., Jessen, A.D., Hergel, I.-M., Clausen, T.D., Larsen, J.E., van Hall, G., Jensen, A.K., Molsted, S., Bendox, J.M., Løkkegaard, E., & Stallknecht, B. (2022). Effects of structured supervised exercise training or motivational counseling on pregnant women’s physical activity level: FitMum randomized controlled trial. Journal of Medical Internet Research, 24(7), Article 37699. https://doi.org/10.2196/37699
Krøner, F.H., Knudsen, S.D.P., Roland, C.B., Alomairah, S.A., & Molsted, S. (2020). Validity and reliability of the Danish version of the pregnancy physical activity questionnaire to assess levels of physical activity during pregnancy. Journal of Maternal-Fetal and Neonatal Medicine, 35(23), 4566–4572. https://doi.org/10.1080/14767058.2020.1856807
Löf, M. (2011). Physical activity pattern and activity energy expenditure in healthy pregnant and non-pregnant Swedish women. European Journal of Clinical Nutrition, 65129(10), 1295–1301. https://doi.org/10.1038/ejcn.2011.129
Mottola, M.F., Davenport, M.H., Ruchat, S.M., Davies, G.A., Poitras, V.J., Gray, C.E., Jaramillo Garcia, A., Barrowman, N., Adamo, K.B., Duggan, M., Barakat, R., Chilibeck, P., Fleming, K., Forte, M., Korolnek, J., Nagpal, T., Slater, L.G., Stirling, D., & Zehr, L. (2018). 2019 Canadian guideline for physical activity throughout pregnancy. British Journal of Sports Medicine, 52(21), 1339–1346. https://doi.org/10.1136/bjsports-2018-100056
Moyer, C., Reoyo, O.R., & May, L. (2016). The influence of prenatal exercise on offspring health: A review. Clinical Medicine Insights: Women’s Health, 9, 37–42. https://doi.org/10.4137/CMWh.s34670
Murakami, H., Kawakami, R., Nakae, S., Yamada, Y., Nakata, Y., Ohkawara, K., Sasai, H., Ishikawa-Takata, K., Tanaka, S., & Miyachi, M. (2019). Accuracy of 12 wearable devices for estimating physical activity energy expenditure using a metabolic chamber and the doubly labeled water method: Validation study. JMIR MHealth and UHealth, 7(8), Article 13938. https://doi.org/10.2196/13938
Murakami, H., Kawakami, R., Nakata, N., Ishikawa-Takata, K., Tanaka, S., & Miyachi, M. (2016). Accuracy of wearable devices for estimating total energy expenditure comparison with metabolic chamber and doubly labeled water method. JAMA Internal Medicine, 176(5), Article 699. https://doi.org/10.1001/jamainternmed.2016.0216
O’Driscoll, R., Turicchi, J., Beaulieu, K., Scott, S., Matu, J., Deighton, K., Finlayson, G., & Stubbs, J. (2020). How well do activity monitors estimate energy expenditure? A systematic review and meta-analysis of the validity of current technologies. British Journal of Sports Medicine, 54(6), 332–340. https://doi.org/10.1136/bjsports-2018-099643
Pedersen, E., Mortensen, L.H., Brage, S., Bjerregaard, A.L., & Aadahl, M. (2018). Criterion Validity of the Physical Activity Scale (PAS2) in Danish adults. Scandinavian Journal of Public Health, 46(7), 726–734. https://doi.org/10.1177/1403494817738470
Perales, M., & Artal, R.L.A. (2017). Exercise during pregnancy. The Journal of the American Medical Association, 317(11), 1113–1114.
R Core Team. (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.r-project.org/
Roland, C.B., Knudsen, S.P., Alomairah, S.A., Andersen, A.D., Bendix, J., Clausen, T.D., Molsted, S., Jensen, A.K., Teilmann, G., Jespersen, A.P., Larsen, J.E., Hall, G.V., Andersen, E., Barrès, R., Mortensen, O.H., Maindal, H.T., Tarnow, L., Løkkegaard, E.C.L., & Stallknecht, B. (2021). Structured supervised exercise training or motivational counselling during pregnancy on physical activity level and health of mother and offspring: FitMum study protocol. BMJ Open, 11(3), Article 43671. https://doi.org/10.1136/bmjopen-2020-043671
Sattler, M.C., Jaunig, J., Watson, E.D., van Poppel, M.N.M., Mokkink, L.B., Terwee, C.B., & Dietz, P. (2018). Physical activity questionnaires for pregnancy: A systematic review of measurement properties. Sports Medicine, 48(10), 2317–2346. https://doi.org/10.1007/s40279-018-0961-x
Schober, P., & Schwarte, L.A. (2018). Correlation coefficients: Appropriate use and interpretation. Anesthesia and Analgesia, 126(5), 1763–1768. https://doi.org/10.1213/ANE.0000000000002864
Schuster, S. (2016). Measuring physical activity in pregnancy using questionnaires: A meta-analysis. Acta Clinica Croatica, 55(3), 440–451. https://doi.org/10.20471/acc.2016.55.03.13
Shei, R.-J., Holder, I.G., Oumsang, A.S., Paris, B.A., & Paris, H.L. (2022). Wearable activity trackers–advanced technology or advanced marketing? European Journal of Applied Physiology, 122(9), 1975–1990. https://doi.org/10.1007/s00421-022-04951-1
Shook, R.P., Yeh, H.W., Welk, G.J., Davis, A.M., & Ries, D. (2022). Commercial devices provide estimates of energy balance with varying degrees of validity in free-living adults. Journal of Nutrition, 152(2), 630–638. https://doi.org/10.1093/jn/nxab317
Siddall, A.G., Powell, S.D., Needham-Beck, S.C., Edwards, V.C., Thompson, J.E.S., Kefyalew, S.S., Singh, P.A., Orford, E.R., Venables, M.C., Jackson, S., Greeves, J.P., Blacker, S.D., & Myers, S.D. (2019). Validity of energy expenditure estimation methods during 10 days of military training. Scandinavian Journal of Medicine and Science in Sports, 29(9), 1313–1321. https://doi.org/10.1111/sms.13488
Speakman, J.R., Yamada, Y., Sagayama, H., Berman, E.S.F., Ainslie, P.N., Andersen, L.F., Anderson, L.J., Arab, L., Baddou, I., Bedu-Addo, K., Blaak, E.E., Blanc, S., Bonomi, A.G., Bouten, C.V.C., Bovet, P., Buchowski, M.S., Butte, N.F., Camps, S.G.J.A., Close, G.L., … Wong, W.W. (2021). A standard calculation methodology for human doubly labeled water studies. Cell Reports Medicine, 2(2), Article 100203. https://doi.org/10.1016/j.xcrm.2021.100203
Strath, S.J., Kaminsky, L.A., Ainsworth, B.E., Ekelund, U., Freedson, P.S., Gary, R.A., Richardson, C.R., Smith, D.T., & Swartz, A.M. (2013). Guide to the assessment of physical activity: Clinical and research applications: A scientific statement from the American Heart association. Circulation, 128(20), 2259–2279. https://doi.org/10.1161/01.cir.0000435708.67487.da
Westerterp, K.R. (2017). Doubly labelled water assessment of energy expenditure: Principle, practice, and promise. European Journal of Applied Physiology, 117(7), 1277–1285. https://doi.org/10.1007/s00421-017-3641-x
Westerterp, K.R., Carter, C., & Yves, S. (2013). Physical activity and physical activity induced energy expenditure in humans: Measurement, determinants, and effects. Frontiers in Physiology, 4, Article 90. https://doi.org/10.3389/fphys.2013.00090
Woolley, S.I., Collins, T., Mitchell, J., & Fredericks, D. (2019). Investigation of wearable health tracker version updates. BMJ Health and Care Informatics, 26(1), Article 100083. https://doi.org/10.1136/bmjhci-2019-100083
Wright, S.P., Hall Brown, T.S., Collier, S.R., & Sandberg, K. (2017). How consumer physical activity monitors could transform human physiology research. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology, 312, R358–R367. https://doi.org/10.1152/ajpregu.00349.2016
